黄河流域极端气候变化气候模式优选

Optimization of Climate Model for Extreme Climate Change in the Yellow River Basin

探讨全球气候变化背景下黄河流域极端气候事件的变化特征,对有效应对极端气候事件及其次生灾害具有重要意义。基于1961—2005年流域历史气象要素指标观测数据,选取10个CMIP5模式,采用逐日偏差校正方法对模拟数据进行偏差较正,并对校正前后的各指标进行对比,从变化趋势和空间分布两个方面对CMIP5模式模拟的数据进行评估,优选出适合黄河流域的气候模式。结果表明:10个CMIP5模式模拟的日最高气温都表现出上升趋势,上升速率在0.18~0.35℃/10 a之间,大多数CMIP5模式模拟的日最低气温上升速率在0.20~0.41℃/10 a之间,MPI、GFDL、CSIRO和CCSM4模式模拟的年降水量表现出下降趋势;各模式模拟的相对误差较小,MPI、MIROC5、GFDL等模式模拟的气温在黄河中下游表现出高估的情况,大多数模式模拟的最大1 d降水量在黄河中游表现出低估的情况;通过数据对比优选出GFDL、FGOALS和CCSM4这3个最适合黄河流域的气候模式。

Discussing the changing characteristics of extreme climate events in the Yellow River Basin under the background of global climate change is of great significance for effectively dealing with extreme climate events and their secondary disasters.Based on the observation data of historical climatic factor indicators in the basin from 1961 to 2005,ten CMIP5 models were selected,and the daily deviation correction method was used for bias correction for the model simulation data,and the indicators before and after the correction were compared.From the change trend and spatial distribution,evaluated the data simulated by the CMIP5 model,and selected climate models suitable for the basin.The results show that the daily maximum air temperature simulated by 10 CMIP5 models all show an upward trend,and the rising rate is between 0.18 and 0.35℃/10 a.The daily minimum temperature rise rate simulated by most CMIP5 models is between 0.20 and 0.41℃/10 a.The annual precipitation simulated by the MPI,GFDL,CSIRO and CCSM4 models shows a downward trend.The relative error of each model simulation is small,the temperature simulated by MPI,MIROC5,GFDL and other models shows an overestimation level in the middle and lower reaches of the Yellow River,and the maximum 1-day precipitation simulated by most models shows an underestimation level in the middle and lower reaches of the Yellow River.Through data comparison,the three most suitable climate models for the Yellow River basin,GFDL,FGOALS and CCSM4 are selected.